They say testing triangle but you are doing microservices. Now what?
summary
Argues that the Unit/Integration/GUI triangle breaks for highly distributed microservices and replaces it with goal‑driven tests: fast unit checks, framework‑aware integration, IO tests against real services, and black‑box Dockerized runs. Demonstrates Spring plus WireMock and Testcontainers running real MySQL/Redis/queue instances, notes Spring context initializes once while containers can start per test (example: 20 tests × ~30 seconds container start becomes ~4 minutes), and shows Docker Compose black‑box tests that mimic production and catch deployment/runtime mismatches so deployments and production experiments succeed.
AI-written summary · about
chapters (18)
- 0:00Intro
- 1:42microservices
- 2:41the testing triangle
- 4:02Jakubs background
- 4:52Problems with testing triangle
- 5:56Unit testing
- 12:40Spring testing
- 15:41Basic integration testing
- 16:43Response time
- 19:02Testing budget
- 24:36Advanced integration testing
- 26:22Quiz
- 35:26Blackbox testing
- 37:21Coding scenarios
- 39:06Summary
- 43:23Experiments on production
- 43:43What you cant test
- 45:42Questions
related talks
47mMicroservices Testing with Java
47mAnother Way to Test Your Microservices
02h 36mTesting Java Microservice
49mEvent Driven Architecture: Love Triangle in Integration Testing
42mMicroservice Testing Techniques
42mUse Docker to test microservices
30mTest as You Run - Production-like Integration Tests with Docker
27mTesting Cloud Apps: Mocks vs. Service Virtualization vs. Remocal Tools